CA1204070A - Biosynthesis of unnatural cephalosporins - Google Patents

Abstract

ABSTRACT OF THE DISCLOSURE
A process for converting a peptide precursor of the ACV type, in which the valine moiety may be replaced by any readily available amino acid,to an unnatural cephalosporin of the type

Description

~" ~2~7Q
)SYNTHESIS OF UNNATURAL CEPHALOSPORINS
C~ .
Field o~ the Invention _ This invention relates to a cell-free process or producing cephalospori~ antibiotics from peptides and derivati~es thereof.

The Prior Art _.
~he be~a-lactam family of na~ural products includes the penicillins:
RCONH ~

N Y

cephalospoxins:
RCONH ~ S

~, CO H
and cephamycins 2 RCONH~ - ~ ~

O

in which the beta-lactam ring i~ fu~ed to a five or six membered sulfur-containing ring; together with cla~ulanic acid OH
~ ~ H
,i J~--~
/~ C02H
in which the beta-lactam is used to a five membered oxyg~n-containing ring;

~2~
~ ~ the carbapenems R ~Rl ~ N~

in which the beta-lactam is fused to a five membered carbon containing ring;
and the R
J~ ~ OH

which are monocyclic compounds.
Although there are many naturally occurring members of this family, only two can be used directly in medicine without structural change. These are penicillin G, the penicillin in which R - benzyl, and clavulanic acid. All other clinically important beta-lactam compounds have been prepared from one or other o~ the natural products by structural change. For many years the changes have been generaly effec~ed by substitution around the peripheries of the various ring systems and not in the ring systems themselves. Since 1974, however, efforts have been concentrated on nwclear modification oE a beta~lactam natural product. Such efforts have generally resulted in complex chemical processes containing upwards of 16 steps with the result that the products are obtained in generally low yield and at extremely high costO Moxalactam~, for example, a third generation cephalosporin, is approximately five times more expensive than cephalothin, a first genera tion cephalosporin; and cephalothin is, in turn, approximately fifty ~imes more expensive than ampicillin, a semi~synthetic penicillin (Drug Topics ~ed Book 1981).
Attention has therefore turned to alternative methods of synthesis, and in particular to microbiological methodsO Cell~free syntheses of penicillins and the related cephalosporins are known in -the art and attention is directed to U.S. Patent No. 4,178,210 issued December 11, 1979 to A.L. Demain et all which teaches conversion only of the D-form, penicillin N, to a cephalosporin compound. In U.S.
Patent No. 4J248~966 issued February 3, 1981, ~.L. Demain et al teach the production of isopenicillin derivatives, in a cell-free system using an extract from Cephalosporiumacremonium, from a tripeptide composed o unsubstituted or ~ substituted D-valine, unsubstituted or substituted L-cysteine, and L-a-aminoadipic acid or its analogs. Freezing of the cell~free extract resulted in inactivation of certain enzyme~ so that conversion did not proceed past the isopenicillin stage.
In ~.S. Patent 4,307,192 issued December 22, 1981, A.L. Demain et al teach the use oE a fre~h (i.e. not frozen) cell-free extract oE C. acremonium so as to preserve the racemase (epimerase) agent or agen~ necessary for the conversion of isopenicillin N to penicillin Nl a necessary intermediate step in the process for conversion of L-aminoadipyl-L-cysteinyl-D-valine (abbreviated to LLD in the reference but hereinafter ACV) via an oxidative cyclization step to isopenicillin N, epimerization to penicillin N and oxidative ring expansion to desacetoxycephalosporin C.

~ ~21~

H SH

CH3 (1) COOH o N
H COOH
5 L-aminoadipyl L-cysteinyl D-valine 1 oxidative cyclization H

2 ~ ~ N ~ ~ S ~ CH3 COOH
~ N ~-~ ~OOH
isopeni`cillin N
¦ epimerize and ring expand H
H2N ~f ~ ~s~
COOH J~

D-aminoadipyl COOH
desacetoxycephalosporin C

The activity of the racemase agent in a cell-free e~tract of C~ acremonium wa~ first recognized by Konomi et al, Biochem.
.
J. Vol. 184, p 427-430, 1979, and confirmed by Baldwin et al, Biochem J. Vnl~ 194~ 649-651, 1981, and Jayatilake et al, Biochem. J. Vol. 194, 649 647, 1981 who also recognized the extreme lahility of the racemase agent so that recovery o the racemase agent per se is belie~ed to be impossible. The lability o the racemase agent is beli-eved to preclude ~se of 7~

cell-free extracts of C. acremonium for hiyh yield commercial production of cephalosporins from peptide precursors~
Since about 1978, 6-aminopenicillanic acid has been prodllced commercially by the deacylation of benzyl penicillin using immobilized penicillin acylase (Proc. lST. European Congress of Biotechnology, Dechema Monographs, Volume 82, 162, 1~78~, an numerous other reactions nave been suggested using immobilized biomaterials such as enzymes (Enzyme Engineering Vol. 6, 198~, Plenum).

It is, therefore, an object of the present invention to provide an integrated cell-free process for producing a cephalosporin compound from a peptide of the general formula H SH E~
H2N ~ ~ N ~ ¦~R2 (4) 1 O J /~CH2 1 COOH O N ~
> H COOH

where Rl is hydrog~n, a lower alkyl or functionalized carboxylic group, and R2 is hydrogen or a lower alkyl group, using stable cell-free extracts from prokaryotic organisms.
It is another objec-t of the present invention to provide immobilized cell-ree extracts from prokaryotic organisms so as to permit continuous production oE cephalo-sporins.
ThPse and other objects oE the i.nvention will be apparent from the following description of the preferred embodiments.
Summary of the Invention It has now been discovered that certain cell~Er~e extracts of prokaryotic organisms such as Str~ptomyces cl~vuligerus, Streptom~ces cattleya and Streptomyces lipmanii, can be separated into three fractions ~y a three stage treat-ment to provide three stable a~d separat~ enzymes:
(a) epimerase (~W approx. 60,000) which may be used, for example, to epimerize isopenicillin N to penicillin N;
Ib) cyclase (MW approx~ 36,500) which may be used, for example, to cyclize ACV to isopenicillin N; and (c) ring expansion enzyme ~MW approx. 29,0003 which may be used, for example, to ring expand penicillin N to desaceto~y-eephalosporin C. It has also been discovered that the threeenzymes may be immobilized on a suitable column material and employed for the continuous production of cephalosporins.
Thus, by one aspect of this invention there is provided a process for producing unnatural cephalosporins ]5 of the formula D-Aad-NH J ~b~ R

~ R

where R1 = H, lower alkyl, or funetionalized carboxylic group and R2 = H or lower alkyl and derivatives thereof, eompxising reaeting a starting material comprising L-a aminoadipyl-L-cysteinyl D-valine and analogs thereof in 25 whieh an amino acid is substituted for the valine moi~ty, with cyclase, epimerase and a ring expansion enzyme isolated from a cell-free extract of a prokaryotie organism for sufficient time and in the presence of suffieient co-factors to produce said cephalosporins.

By another aspec-t oE this invention there is provided a process for isola-ting cyclase, epimerase and a ring expansion enzyme from a cell-free extract of a prokaryotic organism comprising:
(a) precipitating contamina-t:ing pro-teins from said cell-free extract by addition of ammonium sulfate to 40g saturation;
tb) separating precipitated protein from a supexnatant;
(c) adding further ammonium sulfate to 70% saturation to said supernatant thereby precipitating desired said enzymes;
(d) suspending said precipitated enzymes in pll 7 buffer; and ~e) chromatographically separating the desired enzymes from each other.
By yet another aspect of this invention there is provided an immobilized enzyme reagent capable of continuously cyclizing~ epimerizing and ring expanding ACV ancl analogs thereof to desacetoxycephalosporin and the respective analogs thereof, comprising an epimerase having a molecular weight of about 60,000 a cyclase having an MW of about 36,500 and a ring expansion enzyme having a molecular weight of about 29,000, derived from a prokaryotic organism, i~nobilized on a diethylaminoethyltrisacryl chromatographic resin.
Brief Description of the Dra~ n~s Figures la - lf are HPLC chromatographs of reaction mixtures at O mins, 15 mins, 30 mins, 45 mins, 60 mins and 75 mins, respectively.

~2~ 7~
Description of the Preferred EmbodimentS
_ In the follow.ing description, reference will be made particularly to -the conversion of ACV (1) -to desace-toxycephalo-sporin C which isuceful as an antibiotic as such or as a starting compound for the production of cephalosporin antibiotics suah as Cephalexin~. It will be appreciated, however, that -the biochemical techniques of the present invention are equally applicable to other starting materials and it is within the purview of the present invention to substitute the valine 10 moiety in the preferred ACV starting material with any of the readily available amino acids for conversion to the analogous cephalosporins which are useful as antibiotics, or as starting materials for antibiotics such as Ceftizoxime~- Thus, the starting material may be regarded as having the general formula ~4) H SH
2N ~ ~ 2 (4 COOH o ~ N r CH2Rl H COOH
~0 where Rl and R2 are as hereinbefore described.
The amino acids which may be used thus include~

Rl R2 Compound ~ . . _ . .
H CH3 Valine H H a-aminobutyric acid H C2H5 allo isoleucine CH3 CH3 isoleucine COOH H glutamic acid CH2NH-CNH2 H arginine NH

CONH2 H glutamine CH2CH2NH2 H lysine ~ . _ The naturally-occurring beta-lactam compounds are formed as secondary metabolites of both eukaryotic and prokaryotic organisms. Simply statëd, a eukaryote is a higher life ~orm, and it has a more complicated cell struc-ture, which restricts -the types of compound~ that can be synthesized or metabolized. Examples of eukaryotic beta-lac-tam-producing organisms are the fungi Penicillium chrysogenum and _ ~ . A prokaryote, on -the other hand, is a lower, earlierr life form, with a more primitive cell structure, which allows a greater variety of chemical transformations to take place. This suggests, again simply, that prokaryotes are more versatile at organic synthesis than are eukaryotes~ pro~ided that this versatility can be understood and controlledn Examples of prokaryotic beta-lactam~producing organisms are the actinomycetes Stre~tom~ces clavuli~erus/ S. cattleya and -S. lipmanii.

As an illustration oE the differing capabilities '~

of eukaryotic and prokaryotic beta-lactam-producing organisms, P. chryso~enum, a eukaryote, synthesizes ACV and converts this peptide to penicillin as the only stable beta-lac-~am-containing end product. C. acremonium, also a eukaryote, synthesizes the same tripeptide and converts thls peptide sequentially to penicillin and cephalosporin. In contrast, the prokaryote S. clavuligerus synthesizes penicillin, cephalosporin and cephamycin from one amino acid~containin~ precursor and, at the same time~ clavulanic acid, from a different ~ .. . .
10 precursor. The prokaryote S. cattleya synthesizes penicillin and cephalosporin from one precursor and, at the same time, the carbapenem,thienamycin, from a differerlt precursor.
SO clavuligerus, for example, is a well known micro-organism and several strains are available, on an unrestricted 15 basis, from the Northern Regional Research Laboratory, Peoria, Illinois, U.S.A~ under the name NRRL 3S85, among othexs. Other prokaryotic organisms, as described above, are equally freely available. The NRRL 3585 organism must be cultured in a medium and under conditions conducive to 20 the production of ~-lactam compounds, as described in more detàil hereinafter.
There are several methods for cell breakage prior to obtaining a cell-free extract, including French pressure cell,Omnimixer-plastic beads and the preferred sonicatlon.
The preferred treatment comprises sonication for 30 seconds on 48 hour washed cells, followed by centrifugation. The supernatant from this treatment is designated l'crude cell-free extract"~ The crude extract may he separated into three ~2~7~
enzyme fractions in a three stage treatment. In the first stage, contaminating proteins are precipitated by addition of ammonium sulfate to 40~ saturation, and separated from the supernatant by centrifugation or other conventional means.
Addition of more ammonium sulfate to 70~ saturation precipi-tates the desired enzyme activities. The resulting pellet, suspended in pH 7 buffer is termed "salt-precipitated cell-free extract" ~SPCFX~. This SPCFX retains all the desired enzyme activities, and shows reduced baseline contamination in HPLC assays. In the second stage, the epimerase (isopenicillin N ~ pencillin N) (MW 60,000) is cleanly separated from the cyclase (ACV ~ isopenicillin N) (MW 36500) and ring expansion (penicillin N ~ desacetoxycephalosporin C) (MW 20,000) enzymes, by gel filtration chromatography of the SPCFX on, for example, Sephadex~ G~200 (Pharmacia, Sweden).
In the third stage, the cyclase and ring expansion enzymes are separated by ion exchange chromatography on, for example, DEAE Trisacryl resin (sold hy L.K.B., Sweden). A 100-fold puriPication of the cyclase is achiQved in -this manner.
Thus, for the Eirst time three distinct enzyme reagents each having a different enz~natic activity and physical characteristics (e.g. different molecular weights) and which are stable over an extended period of time (of the ox~er of months) under suitable storage conditions of temperature and pH (preferably about ~20C and pH7) have been prepared. The enzymes may be stored and used ~uite separately or may be stored and used as a mixture or immobilzed on a column as required.
Analogous treatment using SPCFX from C. acremonium yields the cyclase and ring expansion enzymes only. As noted above the epimerase is entirely absent due to its e~treme lability.
Following preparation of the three enzyme5~ ACV
dimer or an analog thereof as described above, may be reacted therewith under aerobic conditions, in the presence of the required co-factors such as ferrous ions usually in the form of ferrous sulfate, an antioxidan-t such as ascorbic acid, a reducing agent such as dithlo-threitol (DTT) and a cosubs-trate such as a-ketoglutarate, for sufficient time a-t about 20C and at a suitable pH of about 7 in either ba-tch or continuous mode to produce desacetoxycephalosporin C or an analog thereof.
Example 1 Production of SPCFX
(a) Culture of S. clavuli~erus Streptom~ clavul gerus NRRL 3585 was maintained on a sporulation medium composed of tomato paste, 20g; oat-meal, 20g; agar, 25g, in 1 litre of distilled wat2r, pH 6.8.
Inoculated plates were incubated 7-10 days at 28C.
Spores were scraped off into stPrile distilled water (Sml/
plate) and used to inoculate, 2% v/v, 25ml~125ml flask, seed medium of the :Eollowing composition: glycerol, lOml;
sucrose, 20g; soy flour, 15g; yeast extract, lg; tryptone, 5g; K2HPO~, 0.2g in 1 litre of distilled water/ pH 6.5.
Inoculated seed mediurn was incubated 3 days and used to inoculate, 2% v/v, 100 ml amounts of production medium in 500 ml flasks~ Production medium consisted of soluble starch, lOg; L asparagine, 2g; 3-N-morpholinoprop2ne-sulfonic acid, 21g; MgSO~O7H2O~ 0.6g, K2HPO4, 4.4g; FeSO4.7T-l2O, lmg;
MnC12 4H2O, lmg; ZnSO4.7H2O, lmg; and CaC12.2H2O, 1.3mg 7~

in 1 litre of H20, pH 6.8. Inoculated production medium was incubated 40-48h and the cells were then collected hy filtration andused to prepare cell-free extracts. All incubations were at 27C on a ~yro-tory shaker (250rpm, 19 eccentricity).
(b) Preparation of Cell-Free Extracts Cell-free extracts were ~repared by washin~ 40 48h cells of S. clavul~Lexus in O.05M Tris-HCl buffer, pH 7.0~0.1mM dithiothreitol (DTT) (lOOml/lOOml culture).
10 Washed cells were resuspended to 1/10 of the original culture volume in the same buffer and disrupted by sonication in an ice water bath for 2x15 sec at maximum intensity (300 watts, Biosonik III, Bronwill Scientific). Broken cell suspensions were centrifuged lh at lOO,OGOxg. All cell-free extracts were stored frozen a-t -20C.
Salt~precipitated cell~free extract was prepared by gradual aadition of streptomycin sulfate to cell-free extract with gentle stirring at 4C to a final concentration of 1~, w/v~ Af-ter 15 min at 4C; precipitated nucleic acid 20 wa5 removed by centrifugation for 15 min at 15~000xg. Solid ammonium sulfate ~as then graduall.y added to the supernatant with gentle stirri.ng at 4C until 40% saturation was reached.
After 15 min at 4C ~he suspension was centrifuged as above and the pellet discardedO Additional ammonium sulfate was then adaed to the supernatant, as above, until 70% saturation was reached~ Followlng centrifugation, the pellet was resuspended to its original volume in 0.05M Tris-HCl buffer pH 7.0 containing O.lmM DTT. The enzyme solution was then concentrated to 1/10 of the original volume by ultraEiltration 7~

with an Amicon~PM-10 fil-ter.
Cyclization Assay_~y~
Cyclization activity of enzyme preparations was measured in reaction mixtures containing: bis- 6 ~(L-a-amino-adipyl-L-cysteinyl-D-valine) (ACV)2 0.306mM, DDT 4mM, Na ascorbate 2.8mM, FeS04 45~ M, tris-HCl buffer 0.05M, pH 7.0, enzyme preparation 0003-0.3ml, final volume 004mlO Reaction mixtures were incubated at 20C for up to 4 hour~ and stopped by cooling on ice or by the addition of 0.4ml methanol.
Rin~ Ex~nslon Assay System Ring expansion activity was followed using the cyclization assay system described above but supplemented with ATP 0.5mM, a-ketoglutarate lmM, KCl 7.5mM, and MgS04 7 . 5mM. Total volume and incubation conditions were the same as for the cyclization assay.
Example 2 __ Separation_of Enz~me Fractions (a3 Separation of Epimerase by Gel Filtration Chromatography of SPCFX
2.5ml of SPCFX was applied to a Sephade~ G-200 superEine column (2.5cm x 40cm) which had been equilibrated in 0.05M Tris-HCl buffer pH 7.0 containing O.lmM DTT. The column was eluted with the same buffer and 2.5ml fractions were collected. Fractions were monitored for protein by measuring W absorption at 280nm, and were assayed for cyclase, epimerase and ring expansion activities~ Active fractions were pooled and concen-trated by ultraflltrat.ion usln~ an Amicon~ PM-10 filter.

~2~

(b) Separation of Cyclase and Ring Expansion En~yme by Ion Exchange Chroma-tography of SPCFX
2.5ml of SPCFX was applied to a diethylaminoethyl tDEAE~-Trisacryl~ column (1.6 x 25cm) which had been equili-brated in O.lM Tris-HC1 buEfer pH 7.0 containing O.lmM DTT.
The column was washed with 50ml of the above buffer and then eluted with a linear gradient of 15Gml each of initial starting buffer vs 0.4M Tris~HCl buffer pH 7.0 containing O.lmM DTT.
2.5ml fractions were collected and monitored for protein content by measuring UV-absorption at 280nm. The ring expansion enzyme eluted at about llOmM Tris-chloride, the cyclase eluted at about 150mM Tris-chloride and epimerase at about 175r~M
Tris-chloride. Fractions were also monitored for conductivity and were assayed for cyclization, epimerase and ring expansion activity. Active frac-tions were pooled and concentrated and desalted by ultrafiltration using an Amicon~ PM-10 filter. Use of aTris-chloride gradient is believed to he-tter preserve enzyme activity as compared to the more usual NaCl gradient.
Both separations were performed at 4~C, and the enzyme products were stored at -20C or lower as they were found to lose activi-ty overnight at room temperature.
Example 3 Pre aration of Cell~Free Extr~cts for Immobilization P, . __ . . ................................ .
Cell-free extracts were prepared by washing 40-~8h cells of clavuli~erus in 0.05M Tris-HCl buffer, pH 7.0 +
O.lmM dithiothreitol + O.OlmM ethylenediaminetetracetic acid (EDTA buffer~(lOOml/lOOml culturel~ Washed cells were resus-pended to 1/10 of the original culture volume in EDTAbuffer and disrupted by sonication in an ice water bath for 2x15 sec v~

at maximum intensity (300 watts, Biosonik III, Bronwil:L
Scientific). Broken cell suspensions were centrifuged lh at 100,000x~. All cell-free extracts were stored at -20C.
Salt-precipitated cell-free extract was prepared by gradual addition of streptomycin sulfate to cell-free extract wi~h gentle stirring at 4C to a final concentration of 1%, w/v. After 15 min at 4C, precipitated nucleic acid was removed by centrifugation for 15 min at 15,000xg. Solid ammonium sulfate was then gradually added to the supernatant with gentle stirring at 4C until 40% saturation was reached.
After 15 min at 4C the suspension was centrifuged as above and the pellet discarded. Additional ammonium sulfate was then added to the supernatant, as above, until 70% saturation was reached. Following centrifugation, the pellet was resuspended to its original volume inEDTA buffer. The enzyme solution was then concentrated to 1/10 of the original volume by ultrafiltration with an Amicon~ PM-10 filter.
Imm~bilization of Salt-Prec~itated Cell-Free Extract DEAE-trisacryl resin was loaded into a column 0.4 x 5.8cm (packed bed volume, lml), washed with 3 x 2ml of the sameEDTA buffer, and allowed to drain to dryness by gravity.
One milliliter of thesal-t-precipitated cell-free extract above was applied ~o the column. The eEfluent was collected and reapplied to -the column twice to ensure comple-te enzyme loading. The column was washed with 2 x lml of the sameEDTA
buffer, drained dry and centrifuged for 3 min. at 500xg to remove excess buffer. This immobilized enzyme reactor was stored at 4C when not in use~

~.2~

_a ~
Preparation of ACV and Related Comeounds N-BoC S-trityl-L-cysteine was coup].ed with the benzhydryl ester of D-valine to give a fully protected dipeptide(5).
STr BoCNH ~
(5) - N ~

A 15 minute treatment with anhydrous formic acid at room temperature led to crystalline, partially protected peptide (6).
STr H2N ~ (6) ~ N
O
H C02CHPh2 Conversion to fully protected ACV (7) STr BoCNH ~ ~ NH
O ~ N
C02CHPh2 o H
C02CHPh2 was achieved by coupling peptide (6) with (8 BoCNH ~ C02H (8) C02CHPh2 Deprotectlon of l7) was achieved in two s-tages~

~al removal of the trityl group, with iodine in methanol;
(bl remov~l of all other protecting groups by overnight treatment with formic acid, leading to ACV disulfide (9).
The ACV is best s-tored in this form and may be readily con-verted to ACV (1), as needed, with di-thio-threitol. This synthesis i5 readily adaptable to systematic modifications of the aminoadipyl moiety and compounds sllch as N-acetyl~ACV and its cyclic analog N-acetyl isopenicillin N, may be similarly prepared from N acetyl~L-X-aminoadipic acid alpha benzhydryl ester as the starting material.
Example 5 Preparation of ~ (L-~-aminoadipyl)-L-cysteinyl-D-alloisoleucine ~ACI~

RlNH~ ~ ~ ~ 2 5 This compound was prepared Erom L-~-aminoadipic acid, L-cysteine and D alloisoleucine, as described for the synthesis of the natural cephalosporin precursor ~-(L-a-aminoadipyl)~
L--cysteinyl~D-valine by S. Wolfe and M.G. Jokinen, C~nadian Journal of Chemistry, Volume 57, pages 1388 1396, 1979O
This led, succ~ssively, to the fully protected tripeptide (Rl = t-butoxyca~bonyl, R2 = benzhydryl, R3 = trityl), m.p.
91-93 (ethyl acetate-petroleum ether), Rf 0.54 (methylene chloride-ethyl acetate~ 9:1; yellow with palladium chloride), the detritylated compound (Rl = t-butoxycarbonyl, R2 =
benzhydryl~ R3 = disulfide)~ m.p. 114-116~ Ime~hanol~, Rf 0.76 (methylene chloride-ethyl acetate, 4:1, yellow wi-th palladium chloride), and the completely deprotected compound (Rl ~ R2 = H, R3 = disulfide), Rf = 0.22 (methyl ethyl ketone-water-acetic acid, 4:1:1, purple with ninhydrin), lHmr ~D20) ~: 0.90 (3H, d, 6~z), 0.91 53H, 5, 7Hz), 1.30 (2H, m), 1.73 (2H, br t), 1.88 (2H, br t), 2.01 (lH, m), 2.39 ~2H, br t),

3.00 (lHr q, B, 15H~), 3.16 (lH~ q, 5, l5Hz)~ 3.76 ~lH, t, ÇHz), 4O40 (lH, d, 4Hz), 4.73 (lH, br s). The la-tter compound is converted into the active form (Rl = R2 ~ R3 ~ Hl upon treatment with dithiothreitol.
~e~ ~.
P~paration of_~~ (L-a-aminoadipyl)~ cysteinyl D-a-amin butyrate (ACAb) ~ SR3 RlNH ~ ~ N ~

C02R2 ~ N ~ CH2CH3 H

This compound was prepared, as in Example 5, via the intermediates Rl = t-butoxycarbonyl, R2 = benzhydryl, R3 = trityl: R~ 0.63 (toluene-ethyl acetate, 2:1); Rl = t-butoxycarhonyl, R2 = benzhydryl, R3 = disulfide: Rf 0.48 (toluene-ethyl acetate, 2:1); and Rl = R2 = H; R3 = disulfide:
Rf = 0.1 (methyl ethyl ketone-water-acetic acid, 4:1:1), Hmr (D2O) ~: 0.91 (3H, t, 7.5Hz), 1.59-2.00 (6H, mll 2.4].
(2H, t, 7Hz), 3.97 ~lH, q, 8.5, 14Hz), 3.21 (lHI q, 5, 14Hz)~
3.75 (lH, t, 7Hz), 4~18 (lH, q, 5, 8.5Hz), 4~73 (lH, m).
This last compound is converted into the active form (Rl -R2 = R3 - H) upon treatment with dithiothreltol.

~2~Q~

_ ample 7 Cyclization of ~CV
To 0.4ml of reaction mixture were added O.9mM of ACV
dimer as produced in Example 4~ 500OmM Tris-HCl pH 7.0 buffer and a mixture of the three enzymes as produced in Example 1 from a cell-free extract of S. clavuligerus~ toge~her with 45.0 ~ M ferrous sulfate and 2.8rnM ascorbic acid as op-timized amoun-ts of essen-tial co-factors. DTT was added in e~cess of the amount required to reduce ~CV dimer to ACV monomer.
The reaction was continued for approximately 2 hours at 20C
and then terminated by addition of 0.4ml methanol to precipi-tate protein. It was found, by bioassay and HPLC
proc~dures (described in more detail hereinaf-ter) that the peptide had been convertecl to a mixture of isopenicillin N
and penicillin N. Ring expansion to a cephalosporin did not occ~r. The experiment was repeated with the addition oE ]mM
of a standard oxygenase type enzyme co-factor, alpha-ke-toglutarate, and in this case it was found tha-t the ACV
was converted to desacetoxycephalosporin C.
Example ~
The procedures of Example 7 were repeated using L-aspar-tyl, L-glutamyl, D-a-aminoadipyl, adipyl, glycyl-L-a-aminoadipyl and N-acetyl-L-a-aminodipyl-containing pep-tides.
It was ound that the L-aspartyl, L-glutamyl and D-a-amino-adipyl-containing peptides did not cyclize. Cyclization was observed with adipyl, glycyl-L-a-aminoadipyl and N-acetyl-L-a-aminoadipyl-containing peptides. The adipyl compound gave _ ~0% cyclization to the corresponding penicillin, carboxybutylpenicillin,but ~PCFX con~erted the glycyl and N-acetyl compounds to penicillin N and isopenicillin N, via an initial deacylation of these peptides to ~CV. Purified cyclase from ~. clavuligerus did not cyclize the glycyl-L-a-aminoadipyl-containing peptide. These results sugges-t that the enzymatic conversion of an ACV analog to an unna-tural cephalosporin nucleus requires (i3 a s-L-a aminoadipyl side chain and (ii~ an enzyme system containing the epimerase. A
prokaryotic system is, thereEore, required. Modification of the valinyl moiety, as noted above, has been considered in detail. Substrates modif ied in the valinyl moiety such as:
S~l L-Aad-NH ~ H

N ~CH 2 R 1 where R1 is H, a lower alkyl or functionalized carboxylic group; and R2 is H or a lower alkyl group may be cyclized with carbon-sul:Eur bond formation with retention o configuration at the beta carbon of the valine analog, leading to isopenicillin N analogs of the type:

R
I,-Aad-NH ~ S ~ 2 J~~---N ~ ,~ (10) o C02H

Following epimer.ization to penicillin N analogsof the type:
D-Aad-N ~ C~2Rl ~ N ~ (11) o C02H

21 ~

ring expansion leads to c~phalosporin analogs of -the type:

D-Aad-NH ~ ~ Rl (12~

with transfer of the beta carbon atom attached to C2 of (11) into C2 of the six memb~red ring.
Example_9 The penicillin and cephalosporin-forming abili~y of the immobilized enzyme reactor as prepared in Example 3 was demonstrated using reaction mixtures containing: bis- S ~

(L-~-aminoadipyl)-L-cysteinyl-D-valine IACV)2 0.306mM, dithio-threitol 4mM, Na ascorbate 2.8~1, FeS04 45 M, a-ketoglutarate lmM, KCl 7.5mM, MgS04 7.5mM~ in TDE buffer, final volume 2.Oml.

2ml of the reaction mixture was applied ~o the immobilized enzyme reactor by means of a peristaltic pump operating at 40ml/h. Effluent was collected into a 13xlOOmm test tube ~rom which the original reaction mixture was pumped, and therefore was recycled continuously through the enzyme reactor. The enzyme reactor was operated ak 21~C and 20 ~1 aliquots were removed at 15 minute time intervals for analysis for antibiotic formation. (Table I).
TABLE I
BIOASSAY OF REACTION MIXTVRES

Sample Zone o~ Cephalosporin C
Time Inhibition "equivalents"
(min) ~mm) (~ g~
O O O
.031 19.5 .0~6 ~5 18.5 .062 21~ .136 21.5 .136 ~ 22 -* One microgram of cephalosporin C "equivalen-t" gives a zone of inhibition equal to that produced by 1 ~g of actual cephalosporin C.
Antibiotic levels increased for 60 min. before leveling off.
Since the bioassays were performed in the presence of penicillinase, the antibio-tic activity detected was due to cephalosporin antibiotics only. We show hereinafter tha-t cephalosporins can also arise from ACV via the production of the penicillin intermediates, isopenicillin N and penicillin N. The immobilized enzyme reactor similarly must form cephalosporins by the sequential cycli~ation, epimerization and ring expansion of the ACV peptide substrate.
Analysis of reaction mix-ture time samples by ~IPLC
is shown in Figure l(a-f). With increasing reaction time the ACV peak (13.8-14.26 min) declined while a new peak at 5~2-5.3 min. increased. The new peak is due to a mixture ofisopenicillin N, penicillin N and desacetoxycephalosporin C.
This peak decreases in area gradually Erom 60 min onwards due to ~he further oxi.dation of desacetoxycephalosporin C to desacetylcephalosporin C. Desacetylcephalosporin C has antibiotic activity, so bioassay results remain constant, but this compound elutes with a retention time of 2.2-2.5 min.
under the HPLC conditions used in this study.
Based on these studies we conclude that the immobilized enæyme reac or is converting ACV via a multi step xeaction involving penicillin intermediates into cephalosporin products. Since previous studies have demonstrated that a~
ketoglutarate is absolutely required for the ring expansion of penicillins to cephalosporins, omission of ~-ke-toglutarate from reaction mixtures should stop the reaction at the level of penicillin N~

_ ample :L0 Bioassay of Beta-lactam Compounds Antibiotic in reaction mixtures was estimated by the agar diffusion method. Cyclization reaction mixtures were bio-5 assayed using Micrococcus luteus ~TCC 9341 and Escherichia coli Ess as indicator organisms. Ring expansion reaction mixtures were bioassayed using E. coli Ess as indicator organism in agar plates supplemented wlth penicillinase at 2x105 units/ml.
High Performance Liquid Chromatography (HPLC~
Methanol inactivated reaction mixtures (from Examples 7 and 8) were centrifuged at 12,000xg for 5 min to remove precipitated protein before analysisO Reaction mixtures from Example 9 were examined directly. The chromatographic equipment used was M-6000A pump, UK-6 injector, M-480 variable wave-length director, M 420 data module and Bondapak-C18 column (Rad Pak A in a Z module) as stationary phase. All equipment was from Waters Scientific Co., Mississauga, OntarioO The mobile phase consisted of methanol/0.05M potassium phospha-te buffer, pH 4.0 (5/95). The methanol content of the mobile phase depended upon the particular separation and the source of the material e.g. Examples 7 and 8 or Example 9. A short precolumn (packed with BondapakCl~/Corasil) was used to guard the main column.
UV-absorbing material was detected at 220nm at a sensitivity of 0.02 AUFS.
Example 11 Cyclization and Rin~ Expansion of Unnatural Peptide Substrates The procedure of Example 7 was repeated with ACV
analogs in which valine was replaced by alpha-aminobutyric acid Rl = R2 = H) and allo-isoleucine (Rl = H/ R2 = C2H5), as follows:

7~

(AC-aminobutyrate)2 (ACAB)2 and ~AC-alloisoleuclne)2(ACI)2 were dissolved ln water, neutralized, and lyophillzed in 0.1 and l.Omg amounts. These peptides were then used as substra-tes in cyclization and ring expansion assays as follows- One hundred micrograms of ~ACV)2 from Example 6 was used as substrate in a cyclization and a ring expansion assay system using O.lml of salt-precipitated cell-free extract as enzyme source in each case. (Final concentration of (ACV)2 is 0.306mM). Identical cyclization and ring expansion assays were set up in which 100 ~g (ACAB)2 or l.Omg (ACI)2 replaced the (ACV)2 as substrate and 0.3ml of salt precipitated cell-free extract was used as enzyme source. No subs-trate controls were also prepared. The reaction mixtures were incubated for 2h at 20C. At the end o incuba~ion 20 ~1 amounts of the cyclization reaction mixtures were bioassayed versu~ _ luteus and E. coli Ess; 20 ~1 amounts of the ring expansion reaction mixtures were bioassayed versus E. coli Ess plus and minus penicillinaseO
The remaining re~ction mixtures were then mixed 2Q with an equal volume o~ methanol and centrifuged in preparation for HPLC analysis.
Cyclization and ring expansion reaction mixtures containing (ACI)2 as substrate and also the no subs-trate controls were analysed using a mobile phase of 20~ Methanol/
80~ KH2PO~, 0.05M adjusted to pH 4~0 with H3P04. Twenty microlitre amoun-ts were injected and eluted at a flow rate of 2ml/min.
Cycl,ization and ring expansion reaction mixtures containinq (ACV)2~ (ACI)2 and (ACAB)2 as substrates and also ~ 25 the no substrate controls were -then analysed using a mobile phase o 5~ Methanol/95% KH2PO~ r 0.05M adjusted to pH 4.0 with H3Po4. Twenty m.icrolitre amounts were injected and eluted at a flow rate of 2rnl/min for 5 mln rising to 3ml/min by 7 min and remaining at 3ml/min for the rest of the analysis time.
Results and Discussion Results of biological assays of the reaction mixtures from Examples 7 and 8 are seen in Table 2. Cycliza-tio~ of (ACV)2 re-sults in formation of a bioactive product. The zone size pro duced on E~ coli ~ss agar plates (230Omm) is equivalen-t -to the zone size which a cephalosporin C solution at 29.3~ g/ml would produce. Cyclization of (AcAs)2 produces a bioac-t.ive pro~uct with antibiotic activity equivalent to a 0.9 ~g/ml solution of cephalosporin C against E. coli Ess. Simllarly cyclization of ~ACI)2 produces a bioactive product witn anti.biotic activity equivalent to a 4.85 ~g/ml solution of c~phalosporin C against E. coli Ess. Ring expansion assays containing (ACV)2 result in formation of penicillinase-.insensitive antibiotic which produces a zone size on E~ coliEss ~ penicillinase plates ~22mm) equivalent to a 7.6~g/ml solution of cephalosporin C. Ring expansion assays containing (ACAs)2 do not form penicillinase-insensitive antibiotic nor do they form any antibiotic affect.ing E. coli Ess. Since antibiotic activity was seer- in (ACAB)2~containing cyclization assay systems, this implies one of two ~hings:
1. The additional componen-ts in a ring expansion reaction mixture inhibit cyclization of ACAB~ or 2. Ring expansion assays containing (ACAB)2 produce a cephalosporin which does not affect E coli Ess ~ Rin~ expansion assays cont~ining (ACI)~ form penicillinase insensitive antibiotic which produces a zone size on E. coli Ess + penicillinase plates (12.5mm) equivalent to a O.9 ~g/ml solution of cephalosporin C.
HPLC analysis of cyclization reaction mixtures containing (ACI)2 as substrat~ was carried out with a mobile phase of 20~ methanol/80% Kll2P04, 0-05~ pl~ 4.00 When compared with the no substrate control, (ACI)2 containing reaction mixtures showed a new peak at 2.66 min. Analysis of ring expansion reaction mixtures under the same conditions did not show any new peak because the region around 2.66 min was obscured by UV absorbing m~terial (a-ketoglutarate), present in both the no substrate control and in the test.
When the mobile phase was cihanged to 5% Methanol/
95% ICH2PO4 ~ 0 . 05M pH 4 . 0 r cyclization reaction mixtures containing (ACI~ now showed the new peak to be at 11026 min.
Ring expansion reaction mixtures containing (ACI)2 showed the new peak to be somewhat (~ S0~) recluced in size with a smaller peak running jus~ in fron~ of the main peak.
This is expected since cephalosporins typically run close to, but just in front of, their corresponding penicillin.
Cyclization reaction mixtures containing (ACAB)2 as substrate showed a new peak in the region of 2.33 min.
The corresponding ring expansion reaction mixtures also show their new peak at 2.3 min. Since ring expansion reaction mixtures do not show bioactivity clespite the presence of this new peak, we conclude that the cephalosporin is being formed bu-~ is of lower arltibiotic activity against Eo coli Ess than the corresponding penicillin. Analysis of (ACV) 2 c~ntaining reac~ion mixtures shows that the natural product formed in cyclizati.on reaction mixtures~ a mixture of isopenicillin N
and penicillin N [(iso)penicillin N]~ elutes a~ a retention time of 5.23 min. Ring expansion results in conversion of some of the penicillin to desacetoxy cephalosporin C which runs wlth a retention time of 4.75 min and does not separate from (iso)penicillin N under these conditions.
Based on these studies, it is concluded that salt precipitated cell-free extract from S. clavuli~erus, can cyclize (ACI)2 and (ACABj2 to form penicillins, in addition to being able to cyclize the natural substrate, (ACV)2. The unnatural penicillins so formed have chromatographic characteristics distinct from (iso)penicilli.n N and there is no evidence for production of (iso)penicillin ~ in reaction mixtures containing unnatural peptide substrates.
The same enzyme preparation can cause ring expansion oF the penicillin formed ~rom (ACI)2, resulting in formation of a new cephalosporin.
20 T~ble 2 . . . _ . . . _ .
Zone of Inhibition(mm) ._ . . . _~ . _ _ __ .
Substrate and E. coli E. coli Ess Assay ConditionsM. luteus EsS~ penlcillinase lACV)2 cyclization29.0 28.0 (ACV)2 ring expansion 28.5 22.0 25 (ACA~)2 cyclization8.0 12.5 (ACAB)2 ring expansion 8.0 0 (ACI)2 cyclization13.0 20.0 (ACI)2 ring expansion 20.0 12.5 no substrate cyclization no substrate ring expansion _ + +_ ,,, ~

_ 28 _ xampl~ 12 The proeedure of Example 9 was repeated by passing two reaetioll mixtures each eontaining lmg of ACV ~from Example

4) throu~h a DEAE -trisacryl column t2ml hed vol.) eontaining 2ml of immobilized SPCFX (prepared as in ~xample 3). Each reaetion mixture was eyeled through the column for 1.5 hours at 40 ml per hour. This resulted in approximately 90~ conver-sion of ACV into a mixture of isopenicillin N9 penicillin N, desaeetoxyeephalospol-in C, and desaeetyleephalosporin C as determined by HPLC.

Claims (23)

THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE
PROPERTY OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:

1. A process for producing unnatural cephalosporins of the formula where Rl = H, lower alkyl, or functionalized carboxylic group and R2 = H or lower alkyl and derivatives thereof, comprising reacting a starting material comprising L-a-aminoadipyl-L-cysteinyl-D-valine and analogs thereof in which an amino acid is substituted for the valine moiety, with cyclase, epimerase and a ring expansion enzyme isolated from a cell-free extract of a prokaryotic organism for sufficient time and in the presence of sufficient co~factors to produce said cephalosporins.

2. A process as claimed in claim 1 wherein said prokaryotic organism is selected from the group comprising S. clavuligerus, S~ cattleya and S. lipmanii.

3. A process as claimed in claim 2 wherein said prokaryotic organism is S. clavuligerus.

4. A process as claimed in claim 1, wherein said amino acid is selected from the group consisting of .alpha.-amino butyric acid, allo-isoleucine, isoleucine, glutamic acid, arginine, glutamine and lysine.

5. A process as claimed in claim 1, in which said co-factors include at least one of ferrous sulfate/ ascorbic acid, dithiothreitol and alpha-ketoglutarate.

6. A process as claimed in claim 1 wherein said cyclase, epimerase and ring expansion enzymes are immobilized on a column and said reaction is a continuous reaction.

7. A process as claimed in claim 1 wherein said reaction is a batch process terminated by addition of methanol as to precipitate proteins.

8. A process for isolating cyclase, epimerase and a ring expansion enzyme from a cell-free extract of a prokaryotic organism comprising:
(a) precipitating contaminating proteins from said cell-free extract by addition of ammonium sulfate to 40% saturation;
(b) separating precipitated protein from a supernatant;
(c) adding ammonium sulfate to 70% saturation to said supernatant thereby precipitating desired enzymes;
(d) suspending said precipitated desired enzymes in pH 7 buffer; and (e) chromatographically separating the desired emzymes from each other.

9. A process as claimed in claim 8 wherein said prokaryotic organism is selected from the group comprising S. clavuligerus, S. cattleya and S. lipmanii.

10. A process as claimed in claim 9 wherein said prokaryotic organism is S. clavuligerus.

11. A process as claimed in claim 8, wherein epimerase is separated from the suspension of (d) by gel filtration chromatography.

12. A process as claimed in claim 8, wherein, after separation of epimerase from the suspension of (d), cyclase and said ring expansion enzyme are separated by ion exchange chromatography.

13. A stable epimerase reagent having a molecular weight of about 60,000 capable of epimerizing isopenicillin N and analogs thereof to penicillin N and the respective analogs thereof, derived from a prokaryotic organism.

14. An epimerase reagent as claimed in claim 13 derived from a prokaryotic organism selected from S. clavuli-gerus, S. cattleya and S. lipmanii.

15. An epimerase reagent as claimed in claim 13 derived from S. clavuligerus.

16. A stable cyclase reagent having a molecular weight of about 36,500 capable of cyclizing ACV and analogs thereof to isopenicillin N and the respective analogs thereof, derived from a prokaryotic organism.

17. A cyclase reagent as claimed in claim 16 derived from a prokaryotic organism selected from S. clavuligerus, S. cattleya and S. lipmanii.

18. A cyclase reagent as claimed in claim 16 derived from S. clavuligerus.

19. A stable ring expansion enzyme reagent having a molecular weight of about 29,000 and capable of ring expanding penicillin N and analogs thereof to desacetoxycephalosporin and the respective analogs thereof, derived from a prokaryotic organism.

20. A ring expansion enzyme as claimed in claim 19 derived from a prokaryotic organism selected from S. clavuligerus, S. cattleya and S. lipmanii.

21. A ring expansion enzyme as claimed in claim 19 derived from S. Clavuligerus.

22. A process as claimed in claim 8 wherein said buffer is a l00mM Tris-chloride buffer.

23. A process as claimed in claim 12 wherein said cyclase and said ring expansion enzyme are eluted from a chromatographic column at 150mM Tris-chloride and ll0mM
Tris-chloride respectively.